Most bought lists die the same way. You pay for 3,000 contacts, you load them, you message 300, you get four replies and two of them are "wrong person". The data was not the problem. The list had no idea who it was for.
You can build a prospect list on LinkedIn by hand, from search, in an afternoon a week, and it will outperform anything you buy — because you decided every line of it. Here is the method, including the filters that matter and the discipline that makes them work.
Why bought data performs worse than it looks
A purchased file is a photograph of a database, taken at an unknown moment, by someone who had no opinion about your offer. Three things break.
Job titles drift. People get promoted, change companies, change scope. A list built eighteen months ago is full of people who no longer own the problem you solve.
The selection logic is not yours. The vendor filtered on "Head of" plus "50-200 employees" plus a country. That is not a niche, that is a demographic. You end up messaging people who share a job title and nothing else.
You cannot see the person. No recent post, no job change, no shared group, no mutual contact. So your first message has nothing to hold on to, and you fall back on the generic opener that everyone now recognises on sight. That is the exact failure mode described in why outreach templates stopped working: the message is fine, the targeting was never specific enough to say anything true.
There is also the boring legal point. Buying and loading personal data you did not collect puts you on the hook under GDPR for a lawful basis, a source and a deletion path. Building from LinkedIn search, by hand, for a business proposition you can justify, is a much shorter conversation with yourself.
Narrow the niche until the list writes itself
The hard part of list-building is not the tool. It is deciding who you are not going to contact.
Test your niche like this: can you name the next ten people on your list without opening LinkedIn? If yes, the niche is tight enough. If you can only describe them in categories — "SMB founders", "HR leaders" — it is still too wide, and your search will return 40,000 results you will never work through.
A usable niche has four parts:
- A role that owns the budget or the pain. Not an industry, a role.
- A context that makes the problem live right now: post-fundraise, new site opening, a first management layer being created, a merger, a season.
- A size band narrow enough that your examples apply. 20-80 employees behaves nothing like 300-1,000.
- A geography or language you can actually serve.
Example. Not "executive coaching for leaders". Instead: "operations directors in 80-250 person industrial companies in the north of England who have just taken on a second site." That sentence is a search query. The wide one is not.
The benefit is not only reply rate. A narrow list lets you reuse proof. The thing you learned from client three is relevant to prospect forty. If your list spans six worlds, nothing you learn compounds.
The filters that actually change who you reach
Free LinkedIn search, people tab, is enough to start. Sales Navigator gives you more filters and saved searches; it is worth paying for once you know your niche, not before.
Filters I use in order of usefulness:
Current job title, with variants. Titles are not standard. "Operations Director", "Head of Operations", "COO", "Site Manager" can be the same person. Search each separately rather than trusting one keyword to catch them all. Use quotation marks for exact phrases, and NOT to exclude the noise — "head of operations" NOT recruiter strips out agencies immediately.
Company headcount. The single filter that most improves relevance. Pick one band and stay in it for a month.
Geography. Tighter than you think. A city plus 25 miles beats a country, because it lets you reference local reality and because you might actually meet them.
Years in current position. Under twelve months is gold for anything to do with change: new people have mandates and budgets. Over five years is gold for anything to do with efficiency: they know where the waste is.
Industry — used as an exclusion, not an inclusion. Industry tags are unreliable, but stripping out staffing, consulting and marketing agencies cleans a list fast if you sell to operators.
Past company and school. Underrated. Shared background is a legitimate reason to write, and it is true, which is the whole point.
Groups and event attendees. People in the same professional group have self-selected into a topic. An event attendee list is a niche someone else assembled for you.
Two things to avoid. Do not filter on "open to work" unless you sell to people in transition. And do not use a tool to scrape thousands of profiles into a file — LinkedIn's User Agreement does not allow third-party software or extensions that automate activity on the platform, scraping included, and accounts get restricted for it.
Build the list in passes, not in one sitting
Treat it as a funnel with three passes, because judgement is what makes a hand-built list better than a bought one.
Pass one — raw search. Run one query, one filter set. Work through the results pages and keep anyone plausible. Target 60-80 names. Stop. Do not run five queries at once; you will lose track of which one produced the good ones.
Pass two — the ten-second look. Open each profile. You are checking three things: are they still in that role, is the company still the size you thought, and is there anything specific on the page — a post, a project, a job change, a way they describe their work. Delete anyone with nothing. A silent, empty profile is a bought contact with extra steps.
Pass three — the reason line. For every survivor, write one sentence in your sheet: why this person, this month. "Just posted about recruiting a second shift supervisor." "Company opened a Leeds site in March." "Describes himself as the only ops person for 140 people."
If you cannot write the reason line, remove them. That rule alone will cut your list by a third and double what you can say in the first message. It feeds straight into a better opener — the hook openers worth stealing only work when you have a true, specific first clause to put in front of them.
Keep the sheet simple: name, profile URL, company, size, the reason line, date added, date contacted, outcome. Nothing else. Fancier trackers never get filled in.
Keep the list alive
A list is a living thing, not a file. Three habits.
Refresh weekly, in one block. Ninety minutes, same slot. Add 40-60 names, retire anyone who changed role out of scope. A list you top up weekly never runs dry; a list you build once runs out in six weeks and you panic-buy.
Save the searches that work. When a query produces replies, save it and run it monthly. New people enter the segment constantly — promotions, moves, new companies. Your best source of fresh prospects is the search that already worked.
Let engagement feed the list. People who comment on your posts, or on the posts of people like your clients, are prospects who have already shown interest in the topic. Add them with the reason line "commented on X". This is also where a clear profile pays off, because they will check you before replying — the profile optimisation checklist covers what they look at in the first six seconds.
One honest note on volume. If you later use software to send the invitations and messages — BIZLINKED's Chrome extension does exactly that, sending invitations and AI-written first messages and follow-ups automatically from your own browser and session, inside limits it sets itself — understand what you are choosing. LinkedIn's rules do not allow third-party automation, and any account using it can be restricted, at any volume. No cap removes that risk. Decide with your eyes open.
What to do Monday morning
Block ninety minutes. Write your niche as one sentence containing a role, a context, a size band and a place. Run one search. Collect 60 names. Open each profile and delete anyone you cannot write a reason line for — expect to lose 20. You will finish with roughly 40 people you can name, in a segment you chose, with a true sentence attached to each.
That is a prospect list. It cost you an afternoon and nothing else, and unlike the file you were about to buy, it will still be accurate next week because you will be the one maintaining it.
